As a new iteration of OpenAI’s generative AI technology ChatGPT gains in popularity, what should HR teams and recruiters know about the risks and rewards? Iffi Wahla explains.
It feels like we are living in a time where every week we see a new groundbreaking innovation in artificial intelligence, each more astonishing than the last.
ChatGPT-4o (the “o” stands for omni), the latest iteration of OpenAI’s generative AI, is capable of real-time reasoning across audio, vision, and text. This development marks a significant step towards creating photorealistic AI that can fully mimic human interactions.
With each leap forward in the AI revolution, we face a plethora of questions regarding the benefits and risks of these solutions.
In the recruitment industry, this could help businesses hire top talent from around the globe, but it also brings with it the challenge of how to ascertain whether the person they are considering hiring is genuine, and not someone relying on AI in the application process.
Is it safe?
So what restrictions should be placed on using generative AI in recruitment, and what safeguards can maintain the integrity and quality of the hiring process?
Currently, generative AI can create CVs, cover letters, and applications. It can also assist candidates in researching, preparing answers to common interview questions, and learning best practices in everything from body language to salary negotiation.
Most recruiters and HR professionals have likely encountered applications that appear to have been generated by AI tools like ChatGPT. Generative AI content often stands out due to its elaborate language, odd synonym choices, vagueness, and sometimes, factual inaccuracies.
However, as tools like ChatGPT-4o evolve, distinguishing AI-generated content will become increasingly difficult, and this opens up new avenues for misuse.
Candidates could create a fictitious digital history, such as deepfake videos showing them speaking at events to falsely demonstrate thought leadership or an active blogging presence.
A real-time AI deepfake could even replace a candidate during phone or video interviews. This scenario is not far-fetched – earlier this year, a Hong Kong company lost $26 million to fraudsters who used a deepfake video call to impersonate their chief financial officer and request a money transfer.
Spectrum of support
AI support for candidates exists on a spectrum. Few would begrudge using AI for research – it is essentially a more efficient (if often less accurate) way of googling publicly available information.
However, the use of AI in generating CVs or application answers is more contentious. Some argue that this levels the playing field for individuals who possess the necessary skills but lack the writing abilities or confidence to express them adequately.
Conversely, for positions requiring strong writing skills, this could be a clear misrepresentation.
Generative AI also enables mass and automated applications for roles, raising questions about a candidate’s seriousness if they can quickly apply to dozens of positions simultaneously.
When it comes to the interview process, using AI to tailor perfect responses to potential questions can be seen as merely a more efficient way of preparing answers.
Candidates often already modify their responses to meet the interviewer’s perceived expectations and research the best ways to answer common questions. However, there is a significant difference between rote memorising AI-generated responses and researching and articulating answers in one’s own words.
Hiring someone who can recite AI-generated answers merely proves the candidate has a good memory.
It can also obscure the true abilities of a candidate who may lack confidence and use AI as a crutch. Essentially, AI has the potential to level the playing field to such an extent that differentiating between candidates becomes much harder.
Create guidelines
The issue of generative AI in recruitment is complex and multifaceted. Therefore, it is essential to develop guidelines that govern the recruitment process now.
A straightforward approach is to ban generative AI-created CVs, cover letters and applications unless the position does not require writing skills. Numerous tools can detect AI content, which can be used to screen applications as they are submitted.
Hiring someone who can recite AI-generated answers merely proves the candidate has a good memory
Next, consider the interview process. Increasingly, businesses will use generative AI to create interview questions, which candidates can anticipate using AI.
Effective interviews should avoid generic questions and be tailored to the individual and the role, evolving as the conversation progresses. A dynamic interview is harder to fake through AI preparation and should result in a more accurate assessment of a candidate’s potential and fit.
Interviewing is a skill that can always be improved. Reviewing your processes in light of AI challenges presents an opportunity to refresh procedures and upskill your team.
When it comes to online tasks and assessments, candidates could do written tasks which are entirely AI-generated, or go as far as using a deep fake video to show them doing a test in real time.
Risk vs reward
There are many new ways to cheat the process. The simplest solution is to conduct all assessments in person, although this could be impractical for businesses relying on a global workforce.
The only viable answer is to enhance security procedures to include robust authentication and monitoring, accounting for AI use. The same applies to phone and video interviews to combat AI fakes.
Ultimately, the potential of generative AI to transform recruitment is immense. For every way AI can make recruitment faster, cheaper, and more effective, it also poses challenges in determining whether a candidate is genuinely leveraging new tools or using technology to misrepresent themselves.
Each company and recruiter will need to develop its own policy on where to draw the line, but it should also welcome many of the benefits that it will bring when it comes to hiring global talent, assuming such checks and balances are properly in place.
Sign up to our weekly round-up of HR news and guidance
Receive the Personnel Today Direct e-newsletter every Wednesday
